Académique Documents
Professionnel Documents
Culture Documents
BMJ Open: first published as 10.1136/bmjopen-2013-004461 on 26 March 2014. Downloaded from http://bmjopen.bmj.com/ on 4 July 2019 by guest. Protected by copyright.
Sexual minority population density
and incidence of lung, colorectal and
female breast cancer in California
Ulrike Boehmer,1 Xiaopeng Miao,2 Nancy I Maxwell,1 Al Ozonoff3
BMJ Open: first published as 10.1136/bmjopen-2013-004461 on 26 March 2014. Downloaded from http://bmjopen.bmj.com/ on 4 July 2019 by guest. Protected by copyright.
orientation. To overcome this lack of data on cancer dis- information about the survey methodology can be
parities by sexual orientation, three previous studies obtained from the website: http://www.chis.ucla.edu/.
used county-level ecological analyses to relate breast, The CHIS collects information biennially, including data
lung and colorectal cancer incidence to greater sexual on sexual orientation. To ascertain sexual orientation,
minority population density.21–23 Previously, we used respondents were asked about their sexual identity, with
SEER registry data and US Census data on individuals response choices of heterosexual, LGB along with celi-
living in same-sex partnered households as a proxy bate or other, while recording refusals and do not know
measure for sexual minority orientation. These eco- responses. We combined 4 years of data, using the adult
logical studies concluded that greater female sexual CHIS surveys from 2001, 2003, 2005 and 2007 to
minority density (SMD) in a county is associated with increase the numbers of individuals who report a sexual
greater incidence of breast and colorectal cancer,21 22 minority orientation, defined as gay, lesbian or bisexual.
while there was a negative relationship between female Data on cancer incidence were taken from the
sexual minority population density and lung cancer inci- California Cancer Registry, which records all cancer
dence.23 Greater density of sexual minority men was cases to monitor the occurrence of cancer among
associated with higher incidence of lung and colorectal Californians. We chose the cancer data for the years
cancer.22 23 2001–2008 because these years cover the same time
While US Census data on same-sex households are a frame as the CHIS sexual minority data. We further
well-established proxy for sexual orientation, these data restricted our data to men and women aged 18–84,
also have known limitations. From available Census data, because we focused on adult cancers and also because
we can enumerate households led by same-sex adults, this is the age range for which the CHIS had sexual
but this surely represents an undercount since we orientation data available.
capture only sexual minority individuals who live with a Because there are differences in cancer incidence by
same-sex partner, thereby excluding sexual minority age and race, we used the population data from Census
individuals who are living alone or with non-partners. 2000 since it has age-specific and race-specific popula-
Further, we do not know how many members of a tion data available for all the 58 counties in California.
same-sex partnered household identify as LGB. In par- From the Census, we also obtained county-level data on
ticular, it is impossible to assess lesbians separately from poverty, another important confounder of cancer
bisexual women or to assess gay men separately from incidence.
bisexual men. This is an important limitation, because
studies that analysed lesbian women separately from Measures
bisexual women concluded that bisexuals fare worse on Counts of breast, lung and colorectal cancer incidence
some health indicators than both heterosexual and were classified into 1 of 11 age categories and four race/
lesbian women.24–27 ethnicity groups. The 11 age categories were 18–24, 25–
To address the limitations of US Census data, the 29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64,
present study traded a national scope for an improved 65–69 and 70–84, while race and ethnicity consisted of
measure of SMD. We used statewide population-based (1) non-Hispanic white, (2) Hispanic, (3) Asian/Pacific
data on sexual minority identity (gay, lesbian or bisex- Islander and (4) other race/ethnicity. The group of
ual) to estimate its relationship to colorectal cancer, other race/ethnicity combines non-Hispanic blacks
lung cancer and female breast cancer incidence. (6.79% of cancers), other/unknown race/ethnicity
(0.73% of cancers) and non-Hispanic American Indian
(0.14% of cancers). We calculated each cancer inci-
MATERIALS AND METHODS dence rate using the total female and male population
The Institutional Review Board deemed this study between ages 18 and 84 in each county using the Census
exempt from protocol review. Data for this research data.
project were taken from the California Health Interview We adjusted for poverty level in our analyses, with
Survey (CHIS), the largest state health survey conducted poverty level being defined as the percentage of the
in the USA. The CHIS employs a two-stage geographic- population living under the Federal poverty level, which
ally stratified random-digit-dial sample of households, has been found to be the most consistent, easily inter-
surveying one randomly selected adult from each pretable variable which accurately measures socio-
sampled household. The survey is administered in mul- economic disparities in health outcomes.29 30
tiple languages, resulting in a large multiethnic/multi- Our main independent variable is derived from the
racial sample that accurately represents the California CHIS data on participants’ sexual orientation. We are
population living in households. The CHIS response using these data aggregated at the county level and call
rate shows no significant non-response bias by demo- this aggregate variable ‘SMD’ to express variation in the
graphic characteristics such as age, sex, income, educa- density of sexual minority populations in a county. To
tion or employment status28; however, owing to the make these data age specific, we obtained the distribu-
absence of a sampling frame, non-response by sexual tion of sexual minorities, defined as LGBs, across differ-
orientation has not been evaluated. More detailed ent age groups, and combined this information with the
BMJ Open: first published as 10.1136/bmjopen-2013-004461 on 26 March 2014. Downloaded from http://bmjopen.bmj.com/ on 4 July 2019 by guest. Protected by copyright.
county-level SMD. Specifically, we obtained the weighted for men and because our primary parameter estimates
percentage of gay men in a specific age group (denoted remained essentially unchanged when we fitted the two
as ageweight_gay), using the age information on all gay density measures separately All model selections were
men in the 58 California counties. Then we obtained based on the goodness of fit of the models assessed by
the weighted percentage of all men in the specific age the Akaike Information Criterion (AIC) and residual
group (denoted as ageweight_all), using the age infor- diagnostics plots. SAS PROC GENMOD was used to fit
mation on all participants in the 58 California counties. the models, with the offset term as the logarithm of the
Finally, we obtained the count of all adult gay men US Census age–race-stratified total adult population in
(Ngay) and the count of all adult men (Nall ) in the spe- the county. The incidence rate ratio (IRR) was pre-
cific age group, and computed the age-specific gay sented as the measure of the effect of each predictor,
density as: along with its 95% CI and p value. The validity of the
assumptions and the goodness of fit of the assumed
Ngay ageweight gay models were assessed by residual diagnostic plots and
Age specific gay density ¼
Nall ageweight all goodness-of-fit statistics such as the Deviance statistic. All
analyses were carried out in SAS V.9.2 (SAS Institute Inc,
The age-specific lesbian and bisexual population Cary, North Carolina, USA).
density was computed in a similar way, and we consid-
ered bisexual men and women as distinct categories for
analyses.
Regression diagnostics indicated that the Los Angeles RESULTS
County, white race, age 70–84, data point was a poten- There is considerable variation in the SMD measures by
tially influential point. We refit all regression models county. Across 58 California counties, the lesbian density
excluding this data point to ensure that our findings are measure ranges from 0 to 3.05 (median=0.49;
not predominantly dependent on a single observation. mean=0.66; SD=0.78) and the bisexual density measure
Model fit did not improve substantially with the exclu- from 0 to 4.16 (median=1.24; mean=1.16; SD=0.97). Gay
sion, and more importantly, changes in estimates of asso- and bisexual male density measures respectively range
ciations between the density measures and cancer from 0 to 16.02 (median=0.80; mean=1.10; SD=2.19)
incidence were small and well within the SEs. Therefore, and from 0 to 4.07 (median=0.44; mean=0.60; SD=0.76).
we report all regression results with no data exclusions.
BMJ Open: first published as 10.1136/bmjopen-2013-004461 on 26 March 2014. Downloaded from http://bmjopen.bmj.com/ on 4 July 2019 by guest. Protected by copyright.
Table 2 Multivariate Poisson regression analysis for male colorectal cancer*
Regression
Predictors coefficient estimate IRR 95% CI of IRR p Value
Gay density −0.0020 0.998 0.993 to 1.003 0.45
Bisexual density 0.0262 1.027 1.004 to 1.050 0.02
Hispanic vs non-Hispanic white 0.0411 1.042 1.018 to 1.067 0.0006
Asian/PI vs non-Hispanic white 0.1056 1.111 1.082 to 1.142 <0.0001
Other vs non-Hispanic white 0.0942 1.099 1.065 to 1.134 <0.0001
Poverty 0.0038 1.004 1.002 to 1.006 <0.0001
*Results also adjusted for age.
IRR, incidence rate ratio; PI, Pacific Islander.
Sexual orientation, demographics and cancer incidence significantly associated with decreased breast cancer
among females incidence.
Controlling for race/ethnicity and poverty, each
one-unit increase in lesbian density was associated with a
5.1% decrease in incidence of lung cancer (IRR=0.949, DISCUSSION
p<0.0001; table 3), whereas each one-unit increase in This study’s findings document disparities in the
bisexual density was associated with an 11.3% increase age-adjusted incidence of three cancers across the coun-
(IRR=1.113, p<0.0001). There were significant negative ties of California: disparities that are associated with the
associations between female lung cancer incidence and density of sexual minorities, controlling for race/ethni-
Hispanic, Asian/Pacific Islander and other races. There city and the prevalence of poverty. Our outcome
was no significant association between female lung measure provides effect estimates in the form of the per-
cancer incidence and poverty. centage change in cancer incidence per one-unit
Controlling for race/ethnicity and poverty, each increase in sexual minority population density.
one-unit increase in lesbian density was associated with a Among men, bisexual density (but not gay density)
2.9% decrease in the incidence of colorectal cancer was significantly associated with cancer incidence.
among women (IRR=0.971, p=0.0095; table 4); bisexual Specifically, bisexual density was associated with lower
density was not associated with incidence of colorectal incidence of lung cancer and less strongly with higher
cancer among women. Hispanic ethnicity had a signifi- incidence of colorectal cancer. Among women, lesbian
cant negative association, whereas Asian/Pacific Islander density (but not bisexual density) was significantly asso-
and other race had a significant positive association with ciated with colorectal cancer. Lesbian and bisexual
incidence of colorectal cancer among women. Poverty density was significantly associated with female breast
was not significantly associated with incidence of colo- and lung cancer incidence. However, these two associa-
rectal cancer in women. tions were opposite in direction: lesbian density was asso-
Controlling for race/ethnicity and poverty, each ciated with lower incidence of lung cancer and higher
one-unit increase in lesbian density was significantly asso- incidence of breast cancer; bisexual density was asso-
ciated with a 2.3% increase, and bisexual density with a ciated with higher incidence of lung cancer and lower
3.2% decrease, in incidence of female breast cancer incidence of breast cancer. For each cancer outcome
(lesbian: IRR=1.023, p<0.0001; bisexual: IRR=0.968, that we studied, Los Angeles county, white race and age
p<0.0001; table 5). Hispanic, Asian/Pacific Islander and 70–84 emerged as an influential outlying data point.
other race/ethnicity had significant negative associations Because Los Angeles county is also the most populous
with breast cancer incidence. Poverty was modestly but county in California, we carefully considered issues
BMJ Open: first published as 10.1136/bmjopen-2013-004461 on 26 March 2014. Downloaded from http://bmjopen.bmj.com/ on 4 July 2019 by guest. Protected by copyright.
Table 4 Multivariate Poisson regression analysis for female colorectal cancer*
Regression
Predictors coefficient estimate IRR 95% CI of IRR p Value
Lesbian density −0.0290 0.971 0.950 to 0.993 0.0095
Bisexual density 0.0268 1.027 0.994 to 1.062 0.1105
Hispanic vs non-Hispanic white −0.0670 0.932 0.909 to 0.957 <0.0001
Asian/PI vs non-Hispanic white 0.1133 1.120 1.090 to 1.151 <0.0001
Other vs non-Hispanic white 0.1616 1.175 1.139 to 1.213 <0.0001
Poverty 0.0020 1.002 1.000 to 1.004 0.056
*Results also adjusted for age.
IRR, incidence rate ratio; PI, Pacific Islander.
around the robustness of our results. We conducted a of a positive association for sexual minority men density
sensitivity analysis, which suggests our findings are not and lung cancer using a different dataset.23 For female
overly dependent on this single data point, lending lung cancer, our current findings for lesbian women are
strength to the overall reliability of our findings. consistent with our previous finding of a significant
The differences between results for lesbian/gay negative association between lung cancer and sexual
density and bisexual density are consistent with an minority women density,23 although the present study
increasing understanding among sexual orientation also found a positive association with bisexual density
researchers that—after years of combining lesbian/gay specifically. Finally, findings for breast cancer show some
and bisexual individuals into one group due to small consistency: in the present study, lesbian density has a
sample sizes—differences between these groups come to significant positive association with breast cancer inci-
the forefront once they are separated.8 31–34 These find- dence, confirming our earlier findings,21 although the
ings reflect methodological improvements on our previ- present study also identified a negative association with
ous studies of SMD and cancer incidence21–23: those bisexual density.
analyses used US Census data on whether a respondent A number of factors are likely to contribute to differ-
lived in a same-sex partnered household, whereas data ences between results of earlier Census-based analyses,
for the current analysis rest on self-reported sexual iden- which did not distinguish between gays/lesbians and
tity from the CHIS, as described above, and distinguish bisexuals, and those of the present study, which relied
between gay/lesbian and bisexual respondents. A com- on self-reported sexual identity. The Census data on
parison of findings from our other analyses to those of same-sex partnered households are probably better indi-
this study shows a mix of consistencies and inconsisten- cators of gay and lesbian than bisexual individuals, given
cies (figure 1). the data showing that the majority of partnered gay and
Using SEER data, colorectal cancer incidence was sig- lesbian individuals have a same-sex partner, whereas
nificantly and positively associated with both sexual partnered bisexual individuals are more likely to be in
minority men density and sexual minority women heterosexual relationships.34 The present study’s ability
density22; that finding aligns with this study’s finding for to distinguish lesbian/gay density from bisexual density
male and female bisexual density and colorectal cancer, is an improvement in the ascertainment of sexual minor-
but is inconsistent with the present findings for gay and ity orientation. This most likely accounts for identifying
lesbian density. In the present study, we identified effects that are opposite in direction, as for example,
among men no association between gay density and lesbian density’s association with a 5.1% decrease and
lung cancer, and a negative association between bisexual bisexual density’s 11.3% increase in lung cancer inci-
density and lung cancer; this conflicts with our finding dence. Another factor may be the relative size of the
BMJ Open: first published as 10.1136/bmjopen-2013-004461 on 26 March 2014. Downloaded from http://bmjopen.bmj.com/ on 4 July 2019 by guest. Protected by copyright.
on sexual minority populations’ health and health beha-
viours more broadly. We suggest that the consistency
with which our ecological analyses identified disparities
in cancer incidence is an opportunity for public health
policy interventions, which are larger in scale, consider-
ing county-level programmes, rather than interventions
that focus on individual behaviour change. We hope
additional research can be performed to identify county-
level factors, such as density of healthcare, the equality
Figure 1 Comparing results for different sexual minority
density measures. Not sig, not significant.
of healthcare for sexual minorities, along with other
known cancer prevention behaviours, such as smoking,
that may affect cancer incidence.
subgroups within the LGB adult population: among While increasing attention is being paid to the collec-
women, more identify as bisexual than lesbian, whereas tion of sexual orientation data in the context of state or
among men, more identify as gay.35 Differences in the federal health surveys, from which one can derive differ-
findings of this study compared to the earlier studies are ences in the prevalence of health risk factors, at present,
also attributable to the difference in geographical scope. there is no systematic surveillance of sexual or gender
The present study is limited to cancer incidence in minorities with respect to cancer. Owing to this omis-
California, while the previous SEER-based study was rep- sion, our goal has been to examine the question about
resentative of the USA. Moreover, the earlier studies cancer disparities by sexual orientation using ecological
used a roughly fourfold larger sample (215 counties analyses. Similarly, motivated by a lack of individual-level
compared to the present study’s 58 counties); thus, in data on sexual orientation and cancer, a recent study
the current study, we had lower statistical power to analysed data on women in same sex and opposite sex
detect associations. Both ecological studies failed to relationships, that is, a proxy for sexual orientation, con-
achieve a complete ascertainment of sexual minority cluding that women in same-sex relationships have
status, albeit for different reasons; the Census-based greater risk for breast cancer-related mortality.36 So far,
studies relied on an enumeration of same-sex partnered evidence is accumulating that sexual minorities carry a
individuals, while the present study relied on population disproportionate burden of cancer; therefore, calls are
estimates of sexual identity. Methodological differences growing louder that cancer registries and SEER ought to
aside, there may be real-world differences between collect sexual orientation data to adequately fulfil their
California and the country as a whole in connections mission of monitoring population health, which has to
between SMD and cancer incidence. include sexual minority populations as well.
This study has the inherent limitations of the eco- Contributors UB originated the study and interpreted the findings. XM
logical study design. In particular, because data are at conducted the analyses and helped to interpret the findings. NIM led the
the county level, it is not possible to link sexual minority writing. AO used his statistical expertise to direct the analysis and helped to
status to cancer risk at the level of the individual. Thus, interpret the findings. All authors contributed in significant ways to the final
version of the article by discussing earlier drafts, reviewing and revising the
our study findings clearly describe links between county- manuscript.
level density measures and cancer incidence and should
not be interpreted as evidence of an association between Funding This work was supported by the National Cancer Institute at the
National Institutes of Health (grant number 1 R03 CA153063-01).
individual sexual orientation and cancer incidence. A
finer scale—for example, at the level of the census tract Competing interests None.
—may provide more insight into patterns of SMD and Ethics approval Boston University Institutional Review Board (IRB).
cancer incidence, but it is not yet clear what is the most Provenance and peer review Not commissioned; externally peer reviewed.
appropriate geographic scale for such studies.
Data sharing statement No additional data are available.
Despite these limitations, and some inconsistencies
between the work described here and in our previous Open Access This is an Open Access article distributed in accordance with
the Creative Commons Attribution Non Commercial (CC BY-NC 3.0) license,
analyses,21–23 a particular strength is that these eco-
which permits others to distribute, remix, adapt, build upon this work non-
logical analyses identified the existence of sexual orien- commercially, and license their derivative works on different terms, provided
tation disparities in cancer incidence at the county level, the original work is properly cited and the use is non-commercial. See: http://
while the consideration of gender minority density was creativecommons.org/licenses/by-nc/3.0/
outside the scope of this study. Future studies are
needed to identify ecological causes for the disparity in
cancer incidence, which are most likely complex, pos-
sibly examining county-level factors related to sexual
REFERENCES
minorities’ access to the healthcare system and quality of 1. U.S. Cancer Statistics Working Group. United States Cancer
care delivery. So far, a contextual understanding of Statistics: 1999–2008 incidence and mortality web-based report.
Atlanta, GA: Department of Health and Human Services, Centers for
sexual minorities’ cancer incidence is lacking as well as Disease Control and Prevention, and National Cancer Institute,
knowledge about the county or neighbourhood effects 2012.
BMJ Open: first published as 10.1136/bmjopen-2013-004461 on 26 March 2014. Downloaded from http://bmjopen.bmj.com/ on 4 July 2019 by guest. Protected by copyright.
2. IARC Working Group on the Evaluation of Cancer Preventive 20. Dibble SL, Roberts SA, Nussey B. Comparing breast cancer risk
Strategies. Weight control and physical activity. Lyon, France: IARC between lesbians and their heterosexual sisters. Womens Health
Press, 2001. Issues 2004;14:60–8.
3. Ryan H, Wortley PM, Easton A, et al. Smoking among lesbians, 21. Boehmer U, Ozonoff A, Timm A. County-level association of sexual
gays, and bisexuals: a review of the literature. Am J Prev Med minority density with breast cancer incidence: results from an
2001;21:142–9. ecological study. Sex Res Soc Policy 2011;8:139–45.
4. Tang H, Greenwood GL, Cowling DW, et al. Cigarette smoking 22. Boehmer U, Ozonoff A, Miao X. An ecological analysis of colorectal
among lesbians, gays, and bisexuals: how serious a problem? cancer incidence and mortality: differences by sexual orientation.
(United States). Cancer Causes Control 2004;15:797–803. BMC Cancer 2011;11:400.
5. Burgard SA, Cochran SD, Mays VM. Alcohol and tobacco use 23. Boehmer U, Ozonoff A, Miao X. An ecological approach to examine
patterns among heterosexually and homosexually experienced lung cancer disparities due to sexual orientation. Public Health
California women. Drug Alcohol Depend 2005;77:61–70. 2012;126:605–12.
6. The President’s Cancer Panel. Promoting healthy lifestyles: policy, 24. Steele LS, Ross LE, Dobinson C, et al. Women’s sexual orientation
program, and personal recommendations for reducing cancer risk. and health: results from a Canadian population-based survey.
Annual Report 2006–2007. Bethesda, MD: U.S. Department of Women Health 2009;49:353–67.
Health and Human Services, 2007. 25. Brewster KL, Tillman KH. Sexual orientation and substance use
7. Lee JG, Griffin GK, Melvin CL. Tobacco use among sexual among adolescents and young adults. Am J Public Health
minorities in the USA, 1987 to May 2007: a systematic review. Tob 2012;102:1168–76.
Control 2009;18:275–82. 26. Tjepkema M. Health care use among gay, lesbian and bisexual
8. Duplicate. Adult health behaviors in different age cohorts by sexual Canadians. Health Rep 2008;19:53–64.
orientation. Am J Public Health 2012;102:292–300. 27. Kertzner RM, Meyer IH, Frost DM, et al. Social and psychological
9. Cochran SD, Keenan C, Schober C, et al. Estimates of alcohol use well-being in lesbians, gay men, and bisexuals: the effects of race,
and clinical treatment needs among homosexually active men and gender, age, and sexual identity. Am J Orthopsychiatry
women in the U.S. population. J Consult Clin Psychol 2009;79:500–10.
2000;68:1062–71. 28. Lee S, Brown ER, Grant D, et al. Exploring nonresponse bias in a
10. Cochran SD, Mays VM, Bowen D, et al. Cancer-related risk health survey using neighborhood characteristics. Am J Public
indicators and preventive screening behaviors among lesbians and Health 2009;99:1811–17.
bisexual women. Am J Public Health 2001;91:591–7. 29. Singh G, Miller B, Hankey B, et al. Area socioeconomic variations in
11. Gilman SE, Cochran SD, Mays VM, et al. Risk of psychiatric disorders U.S. cancer incidence, mortality, stage, treatment, and survival,
among individuals reporting same-sex sexual partners in the National 1975–1999. NCI Cancer Surveillance Monograph Series. Bethesda,
Comorbidity Survey. Am J Public Health 2001;91:933–9. MD: National Cancer Institute, 2003.
12. Drabble L, Midanik LT, Trocki K. Reports of alcohol consumption 30. Krieger N, Chen JT, Waterman PD, et al. Painting a truer picture of
and alcohol-related problems among homosexual, bisexual and US socioeconomic and racial/ethnic health inequalities: the Public
heterosexual respondents: results from the 2000 National Alcohol Health Disparities Geocoding Project. Am J Public Health
Survey. J Stud Alcohol 2005;66:111–20. 2005;95:312–23.
13. Valanis BG, Bowen DJ, Bassford T, et al. Sexual orientation and 31. Institute of Medicine. The health of lesbian, gay, bisexual, and
health: comparisons in the women’s health initiative sample. Arch transgender people: building a foundation for better understanding.
Fam Med 2000;9:843–53. Washington, DC: The National Academies Press, 2011.
14. Case P, Bryn Austin S, Hunter DJ, et al. Sexual orientation, health 32. Bauer GR, Jairam JA. Are lesbians really women who have sex with
risk factors, and physical functioning in the Nurses’ Health Study II. women (WSW)? Methodological concerns in measuring sexual
J Womens Health 2004;13:1033–47. orientation in health research. Women Health 2008;48:383–408.
15. Boehmer U, Bowen DJ. Examining factors linked to overweight and 33. Fredriksen-Goldsen KI, Kim H-J, Barkan SE, et al. Disparities in
obesity in women of different sexual orientations. Prev Med health-related quality of life: a comparison of lesbians and bisexual
2009;48:357–61. women. Am J Public Health 2010;100:2255–61.
16. Boehmer U, Bowen DJ, Bauer GR. Overweight and obesity in 34. Herek G, Norton A, Allen T, et al. Demographic, psychological, and
sexual minority women: evidence from population-based data. Am J social characteristics of self-identified lesbian, gay, and bisexual
Public Health 2007;97:1134–40. adults in a US probability sample. Sex Res Soc Policy
17. Deputy NP, Boehmer U. Determinants of body weight among men of 2010;7:176–200.
different sexual orientation. Prev Med 2010;51:129–31. 35. Greenberg JA. Health care issues affecting people with an intersex
18. Rankow EJ, Tessaro I. Mammography and risk factors for breast condition or DSD: sex or disability discrimination? Loyola Los
cancer in lesbian and bisexual women. Am J Health Behav Angeles Law Rev 2012;45:849.
1998;22:403–10. 36. Cochran SD, Mays VM. Risk of breast cancer mortality among
19. Kavanaugh-Lynch MHE, White E, Daling JR, et al. Correlates of women cohabiting with same sex partners: findings from the
lesbian sexual orientation and the risk of breast cancer. J Gay National Health Interview Survey, 1997-2003. J Womens Health
Lesbian Med Assoc 2002;6:91–5. 2012;21:528–33.